comfyui

HiDream-O1 has text encoder refined.

#7
by Spawn - opened

https://huggingface.co/HiDream-ai/Prompt-Refine
This is refine model, and they said more good? or finetuned.
Who can do the hard work so that the tuned model can be used as well?

https://huggingface.co/HiDream-ai/Prompt-Refine
This is refine model, and they said more good? or finetuned.
Who can do the hard work so that the tuned model can be used as well?

https://huggingface.co/tuolaku/Prompt-Refine-GGUF/tree/main

https://huggingface.co/HiDream-ai/Prompt-Refine
This is refine model, and they said more good? or finetuned.
Who can do the hard work so that the tuned model can be used as well?

https://huggingface.co/tuolaku/Prompt-Refine-GGUF/tree/main

HiDream-O1 was checkpoint so self textencoder and self vae used. Now this gguf give same self TE self VAE?
Edit: I knew it is change only TE to this? Okey! Thanks.

https://huggingface.co/HiDream-ai/Prompt-Refine
This is refine model, and they said more good? or finetuned.
Who can do the hard work so that the tuned model can be used as well?

https://huggingface.co/tuolaku/Prompt-Refine-GGUF/tree/main

HiDream-O1 was checkpoint so self textencoder and self vae used. Now this gguf give same self TE self VAE?

Their Prompt-ReFine model is designed solely for prompt enhancement. It is built upon a fine-tuned Gemma4-31B and is intended to be used alongside their provided system prompt.

I had never seen a prompt refiner has 31B parameters, such number is unreasonably large.

Sign up or log in to comment